Currently Online 
| JWeinCom said: Anyway... On polling news. Same old same old, but a few tidbits. Monmouth University has Biden up by an average of 4.5 in Arizona. Monmouth hasn't been as positive towards Biden so that's a decent figures. What's worth noting is that their study is based on two different turnout models. One showing Biden up 2 and one up 7. Two polls out in Virginia show Biden up by 13 and 15. Biden is projected to win by 13.6%. In 2016 Clinton was up by 5 did in fact win by about 5%. In Colorado, Clinton was expected to win by 4 and won by 5. Biden is up by 11.4%. This speaks to the "polls were off in to 2016 argue". In Colorado we're seeing a swing of 8 points, and about 7%. So then, based on only that, what should we expect in states like Wisconsin, Pennsylvania, and Michigan, traditionally purple states where Trump won by about 1%? Well, we'd expect Biden to lead by about 4-6%. What we see is Biden up by 6 in Pennsylvania, 7.6 in Michigan, and 7 in Wisconsin. This is also what we should expect with a national swing about 8 points. How to explain these numbers? Two possible explanations. a) Numbers in the rust belt are still just as off as they were in 2016. In addition, the national polls, polls in Colorado and Virginia, and other states like Texas, South Carolina, Georgia, North Carolina, etc. have also gotten about as bad as the rust belt polls in 2016. or b) The national polls, and polls in those states are still about as good as they were in 2016, and the errors in the few states that were way off have been corrected. Second seems more likely to me, but believe what you will. |
One also needs to consider some states are more or less elastic than others, in response to election swings, and almost all of them are continuously trending blue or red at varying speeds. You also have places like Manhattan or Washington DC where an 8 point shift towards Biden would be mathematically impossible, so the swing must be happening somewhere else.
